Conclusions

In this Chapter, we have presented a study of the impact on video quality of some of the most important parameters supposed to have an impact on it, namely, the stream bit rate, the scene frame rate, the network loss rate, the burst loss size and the ratio of the encoded intra to inter macro-blocs for H263 codecs. Similarly, we have presented in this Chapter a study of the impact of the most important parameters on speech quality, namely the codec used (PCM, GSM and ADPCM), the loss rate, the loss distribution and the packetization interval. As far as we know, there is no previous study of the combined effect of several parameters on neither speech nor video quality. There are several studies in the literature, but they either concentrate on only one or two parameters, and/or they based they study on objective measurements like PSNR or SNR (which do not correlate well with subjective results). The goal of this analysis is to help in the understanding of the behavior of real-time media streams transmitted over best-effort networks. This may be used, for instance, in developing control mechanisms allowing the delivery of the best possible video quality given the current network state. These studies is based on our methodology, presented in Chapter 4 and the results of Chapters 5 and 6. We used the RNN for all our studies given in this Chapter because it behaves better than the standard ANN model. ChapterChapter